Today, automated reasoning is a real need for intelligent systems. Information Retrieval systems in general and specifically a question answering system require a reasoning mechanism as well. In this paper, a heuristic reasoning mechanism, implemented in the online phase of TeLQAS is proposed. TeLQAS is an ontology-based natural language question answering system for the domain of Telecommunications. In an online process, TeLQAS accepts the users’ questions in English. Then, as a first try, the system attempts to obtain reasonable answers based upon the ontology through reasoning. However, the next step is a summarization-based answer extraction, which is not the focus of this paper. The reasoning mechanism utilizes a number of rules defined based on the existing relation types in the ontology and also some linguistics and structural clues of the English Language. The results of experiments prove a considerable improvement in the system performance in comparison with the case where there is no reasoning according to the ontology. In addition to the accuracy of answers, a great reduction in turn-around time is also gained.
[1]
Guy Lapalme,et al.
Goal-Driven Answer Extraction
,
2000,
TREC.
[2]
Kevin D. Ashley,et al.
Reasoning with Reasons in Case-Based Comparisons
,
1995,
ICCBR.
[3]
Grace Hui Yang,et al.
Modeling Web Knowledge for Answering Event-based Questions
,
2003,
WWW.
[4]
Edward Courtney,et al.
2 = 4 M
,
1993
.
[5]
Peter A. Flach,et al.
Learning in Clausal Logic: A Perspective on Inductive Logic Programming
,
2002,
Computational Logic: Logic Programming and Beyond.
[6]
Leila Kosseim,et al.
QUANTUM: A Function-Based Question Answering System
,
2002,
Canadian Conference on AI.
[7]
Elizabeth D. Liddy,et al.
Question Answering: CNLP at the TREC 2002 Question Answering Track
,
2002,
TREC.
[8]
Ryszard S. Michalski,et al.
The Logic of Plausible Reasoning: A Core Theory
,
1989,
Cogn. Sci..